Theta burst stimulation add on to dialectical behavioral therapy in borderline-personality-disorder: methods and design of a randomized, single-blind, placebo-controlled pilot trial

Patients with BPD undergoing an 8–12-week routine inpatient DBT treatment in one specialized ward of a German psychiatric hospital are screened for eligibility after 4 weeks of DBT-treatment. After reaching informed consent they are additionally randomly allocated (1:1) to receive 20 sessions iTBS (DBT + iTBS) or 20 sessions placebo/sham stimulation (DBT + sham; each 1 per day from Monday to Friday) from the 5th to the 8th week of the treatment (see Fig. 1). Randomization procedure is based on an algorithm created with the programming platform MATLAB. Patients are blind regarding intervention (iTBS) or placebo (sham) whereas study personnel is not. Furthermore, data analyses will also be done blinded regarding group allocation.

The study protocol was finished in July 2019 and the trial was registered at DRKS in January 2020. Screening and inclusion of patients started in January 2021. All included patients sign the written declaration of consent, patient information and information on data protection. The study was approved by the Ethics Committee of the medical faculty of the Heinrich Heine University, Duesseldorf.

Routine DBT-treatment is provided based on an adaptation for inpatient treatment in Germany by Bohus et al. [33]. The structured, modularized program contains the modules skills training, interpersonal skills, dealing with feelings, and mindfulness. It is conducted in weekly group sessions including psychoeducation and training elements and additional one to two individual sessions per week. Psychoeducation is aimed at explaining BPD to affected individuals. This includes theories of how the disorder arises, comorbidities and ways to lead a healthy life despite persisting symptoms. Additionally, concepts like the dialectical view on the disorder and the importance of sleep hygiene are taught. Regarding DBT skills, the module aims at teaching simple models of emotion, including strategies on how to control them. Additionally, patients experience ways to deal with interpersonal problems as well as skills to tolerate distress caused by crisis leading to, e.g., self-harm pressure or severe suicidal ideation. The module DBT tools teaches a deeper understanding of oneself, including behavior analysis, problem solving as well as instrumental, primary and secondary emotions. Furthermore, it aims at creating self-compassion in participants. The therapeutic team consists of either fully trained psychotherapists or psychologists in training. The therapist in charge is DBT-certified. Lastly, patients receive psychiatric care by an individually assigned nurse once a week.

Patients between 18 and 45 years of age and diagnosed with a borderline personality disorder and comorbid Major Depression are eligible for our study. Diagnoses are assessed using the standardized clinical interviews, Diagnostisches Kurzinterview bei psychischen Störungen (Mini-DIPS OA; [34]) and the Structured Clinical Interview for DSM-5 Personality Disorders (SCID-V-PD; [35]). All patients participate in an in-patient standardized DBT treatment in our clinic. Patients had to have no previous knowledge of DBT. They need to have a sufficient knowledge of the German language and be able to give informed consent. In case of drug treatment, stable intake in therapeutic doses for two weeks before the start of the study stimulation phase, is necessary, remaining stable during the 4-weeks stimulation phase. For female patients, further inclusion criteria includes negative pregnancy test and willingness to use contraception for the duration of the stimulation. Exclusion criteria are: history of seizures (epilepsy), metallic foreign objects in the skull, pronounced tattoos in the region of the head, significant brain malformations or tumors, cerebral-vascular events, traumatic brain injuries, neurodegenerative diseases, brain surgery, deep brain stimulation, other intracranial implants, cardiac pacemaker, other serious physical illness and other psychiatric comorbidities beside major depression and borderline personality disorder. Moreover, patients with acute suicidality (MADRS score > 4 for question 10), tinnitus, pregnancy, claustrophobia, current or previous treatment with electroconvulsive therapy or vagus nerve stimulation are not eligible. Patients taking anti-epileptic medication including benzodiazepines at a dose equal to or greater than 1 mg/day of lorazepam are also excluded as well as, patients under legal guardianship with reservation of consent.

Before study enrollment, all patients will be scanned to exclude alterations in brain morphology. In case of more than four missed iTBS/sham sessions patients will also be excluded from the analyses. However, all other patients will be included irrespective of later dropout.

The stimulation is applied with a PowerMAG Research 100 magnetic stimulator [36] and a PMD70-pCool figure of eight coil. The devices have the European Certification (CE) mark and are used exclusively for the purpose specified by the manufacturer and are only operated, used and maintained by people who have the necessary training, knowledge and experience. The resting motor threshold will be determined automatically using electromyography before the first treatment and two weeks later, with the integration of the motor evoked potential and the algorithmic determination of the threshold [37‐39]. The stimulation targets the left DLPFC, which will be located using the Beam-F3 method [40]. Each treatment session with iTBS consists of 600 stimuli. The intervention will be applied at 80% of the resting motor threshold intensity. The stimulation is administered intermittently (two seconds stimulation, eight seconds pause), the duration of a treatment unit will be three minutes and twelve seconds. The sham stimulation is carried out with a double coil PMD70-pCool-Sham [41]. The PMD70-pCool-Sham coil's reduced magnetic field strength allows for stimulation of only the nearby scalp, producing a twitching sensation without affecting the brain, and it has equivalent weight and sound to the active coil [41]. Subjects in the sham group experience noises and physical sensations similar to the verum stimulation.

After successful study enrollment, before commencement of the stimulation, the following target parameters are recorded at baseline (see Table 1).

Short version of Borderline Symptom List (BSL-23) [42], second edition of Beck’s Depression Inventory (BDI-II) [43], Global Assessment of Functioning Scale (GAF) [44] and the Baratt Impulsiveness Scale, short form (BIS-15) [45]. Further, the German short version of the Self Compassion Scale (SCS-D) [46], the Montgomery–Åsberg Depression Rating Scale (MADRS) [47] and a Delay Discounting task [48] will be used (see detailed description below).

At the end of each week with iTBS/sham intervention, patients are asked to complete a BSL-23 and BDI-II questionnaire to assess symptom severity. Possible undesirable side effects of the stimulation are assessed and recorded by the practitioner after each treatment according to routine trial (serious) adverse events monitoring. At the end of the intervention period (after four weeks), further assessments are made for MADRS, GAF, BIS-15, SCS-D. Moreover, to check for a potential experimenter bias, participants’ belief about whether they took part in the verum or sham treatment condition are assessed after the first iTBS session and at post-intervention.

Primary outcome is the difference of improvement of borderline symptomatology between the two groups quantified by changes in the short version of the BSL-23 [42, 49] from baseline (T0) to post intervention (T1). The BSL-23 is a validated self-rating measure of borderline personality disorder symptoms. It contains 23 items that are scored from 0 = not at all to 4 = very strong. All items concern the last week, e.g. ‘In the course of last week, I felt helpless’ or ‘In the course of last week I was afraid of losing control’. The BSL-23 has shown good convergent validity, r = 0.89 [49] and reliability, α = 0.94–0.96 [42].

As a secondary target criterion, the decrease in depression is recorded by the MADRS [47] and the second edition of BDI-II [43].

The MADRS [47] is a measure of change in depression. It contains 10 items which are scored from 0 to 6. Each item has its specific meaning for the rating, e.g. the item ‘Inability to feel’ is scored from 0 = “Normal interest in the surroundings and other people” to 6 = “The experience of being emotionally paralyzed, inability to feel anger, grief or pleasure and a complete or even painful failure to feel for close relatives and friends”. Another example item is reduced appetite, scored from 0 = “Normal or increased appetite” to 6 = “Needs persuasion to eat at all”. The MADRS has shown good validity, r = 0.63 and reliability, α = 0.85 [50].

The BDI-II [43] is a widely used scale of depression symptoms. It contains 21 self-report items which are scored from 0 to 3, with 3 being the highest severity. Example items include Crying, scored from 0 = “I don’t cry more often than I used to” to 3 = “I feel like crying, but I can’t” and Worthlessness, scored from 0 = “I do not feel I am worthless” to 3 = “I feel utterly worthless”. The BDI-II has been evaluated as valid d = 0.80 and reliable, α = 0.89 [51].

Improvement in general functional level in the verum vs. sham group is measured by the GAF [44]. The GAF describes an individual’s general functioning on a broad spectrum of activities in 10 steps, which each contain 10%, meaning an individual can score the lowest functional level of 1–10% meaning a persistent danger to self or others, a persistent inability to care for yourself or performed a serious suicidal act. The highest achievable level, 91–100%, no symptoms present, excellent functioning in a wide range of activities as well as good social standing due to self-presentation and a good handling of life obstacles. The GAF has been found to be reliable, r = 0.54 to 0.66 [52] and showing good discriminant validity, r = 0.63 [53].

A decrease in impulsivity is recorded by the BIS-15 [45]. The BIS-15 is a measure used for the identification of impulsive acts. It contains 15 items scored from 1 = ”Rarely/Never” to 4 = “ Almost Always/Always”. Example items include “I say things without thinking” and “I don’t pay attention”. The BIS-15 has been found to be valid, r = 0.37 = and reliable α = 0.81 [54].

Changes in self-compassion will be assessed by the German short version of the SCS-D [46]. The SCS-D measures the ability to have compassion for oneself rather than self-judgement. It contains 12 items scored from 1 = “very rarely” to 5 = “very often”. Example items include statements like “When I feel down, I tend to feel like most other people are probably happier than I am” and “I am disapproving and judgmental about my own flaws and inadequacies”. The short form of the SCS has been shown to be reliable, α = 0.86 and valid, r > 0.55 for all subscales [55].

Changes in impulsivity will be assessed by a cognitive task (delay discounting).

Delay discounting as a representative characteristic of impulsive behavior is especially important in BPD, which is often described as an impulse control disorder [56]. Krause-Utz et al. [57] found that delay discounting might be considered a general feature of BPD, with individuals affected by it showing a higher degree of delay discounting even when controlling for comorbid ADHD. The researchers also concluded that the number of induced stressors did not influence the degree of delay discounting in BPD individuals. This means that the degree of delay discounting is likely not influenced by external factors, but generally present in affected individuals.

The delay discounting paradigm is a very common way of measuring impulsiveness and self-control. The variant we are going to use was developed by Kirby et al. [48] and has already been validated in numerous clinical populations (e.g. in addiction disorders [48]; depression [58]; and schizophrenia [59]). In the assignment, patients will be asked to make a series of decisions where they have to choose between a small reward on the same day or a larger reward that is paid out after a waiting period. The task will include 27 rounds with different amounts of money (11–85 €) and different waiting times (7–186 days). At the end of the experiment, one of these 27 rounds will be randomly selected and the patients will be paid 10% of the chosen amount of money. If the patients have opted for the immediate reward in the selected round, they will receive the money on the same day. If a decision is made in favor of the later reward, the money will either be sent by post or, if the time is still during the inpatient stay, the patients will receive the money from the ward staff on the specified day. Since the patients complete the task twice, they can earn between 2.20 and 17 €.

Structural MRI data will be analyzed using the FreeSurfer software (versions 7.3.2, [60]). Analysis and quality-control protocols of the ENIGMA consortium will applied (http://​enigma.​ini.​usc.​edu/​protocols/​imaging-protocols) covering the recon-all -all stream and cortical as well as subcortical segmentation based on the Desikan–Killiany atlas. Data will be visually inspected and statistically evaluated for outliers. After quality control, volumes of 54 subcortical and cortical regions (27 per hemisphere) will be extracted, in addition to the intracranial volume to correct for global brain volume in statistical models.

In two small studies that examined the efficacy of rTMS in the treatment of borderline personality disorder, medium to large effect sizes were demonstrated (Cohen’s d = 0.74–2.79, [18, 61]). Since these pilot studies had small sample sizes, the possibility of an alpha error and an overestimation of the effect size cannot be ruled out. In order to use a conservative estimate, Whitehead et al. [62] recommend a group size of N = 15 for pilot studies if mean effect sizes are to be expected in the outcome parameter. Since we conduct a pilot study with an inter-subject design with two groups and allow for a 25% contingency for dropouts, we will include N = 53 participants in the study totally for both interventions. Based on a power/sample size calculation with G-POWER [63], the resulting total sample size of 40 patients is sufficient to detect a small to medium effect (f = 0.2) with alpha = 0.05 and a power (1-beta) of 80% (for testing the interaction in a repeated measurement design with two groups). Thus, with the projected sample size a clinical relevant effect of rTMS (in addition to DBT) can be revealed with sufficient power.

Regarding therapy response, a 2 × 2-factorial between-subjects design with the between-subject factor stimulation (TMS vs. Sham) and the within-subject factor time (T0 vs. T1) will be conducted. To deal with a possible bias by treatment drop-out, a Linear Mixed Repeated Measure Model approach will be applied considering all assessed timepoints and the primary outcome (differences in BSL-23 reduction over time between study groups, i.e. interaction time*group) will be tested for significance (p ≤ 0.05). In case of relevant pre-treatment differences, these will be included in the model as covariates to adjust for. Likewise, secondary outcomes will analyzed accordingly.

In a second step, patients will be grouped into responder and non-responder. Group assignment will be performed post-hoc and based on the Reliable Change Index (RCI) as suggested by Jacobson and Truax [64]. An RCI above the 95% confidence limits (+/= 1.96) indicates a reliable change (p < 0.05). To identify parameters which are relevant for therapy response, group comparisons will be performed on secondary outcome variables of T0, using MANOVA models with the independent factor therapy response (responder vs. non-responder) and secondary outcome variables of T0 as dependent variables. Analyses will be corrected for multiple comparisons using Bonferroni Correction and partial eta-square will be determined to report effect size.

Analyses will be conducted with IBM SPSS [65].